with some reasonably accurate insights for you (in fact I'm certain it's already used that way), but using it in this way, where AI is taking the poll? That seems pretty silly.

At least, for now

Hopefully this type of thing is just for research purposes and nobody plans on releasing them publicly, unless it's made very clear to any reader that the results are completely made up by AI. With a clear disclaimer of 'for entertainment purposes only'.

LLM isn't even good at counting. It would suck at statistics.

It would perform the calculations flawlessly. I don't where you get the idea that computers don't 'count' well. In a very pertinent sense, 'counting' is literally 'all a computer does'

As someone who is immersed in AI systems daily, large language models really do suck at math. They first have to recognize that they are in a mathematical context, then they typically OFFLOAD those computations to an external processor, because the way that latent spaces work, they are simply not well-optimized for arithmetic. Now, higher order symbolic mathematics? Yeah, they actually do pretty well with that, because those are essentially manipulation of blocks of symbols (most of which have been tokenized) and pattern matching. However, anyone using LLMs for statistical work is essentially getting the benefit of a lot of back-end python code that was custom written by humans to take those computations out of the equation.

The people that program it...If the programmers have a right wing spin, it will devolve with articulate opinions from the right...And so on.

The #1 thing the AI companies are working on now is training their models to convincingly promote lies and propaganda. Because there is enormous money to be made selling AI with that kind of capability to oppressive governments and corporations.